canon cmos mean time failure

[thomas_sullivan]

ok, searched photo net and googled the web.........nothing (in the

first 10 pages anyhow....get tons of modular transfer rates, but no

mean time failure for MTF)

 

so the question is..........When is the industry saying these chips

in the cameras are gonna fail?........or best guess........or

MTF...........or even your own experience. I suspect it is in the

10's of thousands of exposures range, but depending on how much one

shoots, this could still be relatively short period of time. By

short I mean 5 years. $1500 for a 20D type cam every 5 years is not

my idea of reliability, by the way.......10 years, maybe.

 

This by no means is a film/digital rant. I like the idea of

digital. Just dont like the idea of being in a position to have to

replace the cam every 5 years........hell, I'm just gettin use the

thing by then ;o)

 

any help........links........personal experience.......would be

welcome.

[digitmstr]

personal experience? The Canon CMOS hasn't been around long enough for anyone to give you "personal experience" figures. But, the ones you suggest are quite off mark.

[thomas_sullivan]

"...the ones you suggest are quite off mark....".........based on what? I could see my self shooting 25000+ exposures in 5 years....heck, I shoot roughly 3600 a year using film.

[willie_ju1]

What you want to search for is MTBF - mean time between failures.

[thomas_sullivan]

heh.........thanks willie, that helped tremendously...still havent found an ACTUAL report from Canon or the chip makers, but at least some of the user reports seem hopeful........hmmmmmm, maybe.

 

Although, now, I guess my question is slightly modified. How many have experienced the "hot pixel" problem. That would drive me nuts...one of the reasons I want digital is to get away from having to "dust removal (ie clone)" so much.

[mike sisk]

One possible answer here: <a href="http://www.logicdevices.com/support/

qamanual.PDF">www.logicdevices.com/support/qamanual.PDF</a>

 

<blockquote>"SEMICONDUCTOR FAILURE PATTERNS<br><br>

 

Characterization of semiconductors show that failure rates change during a device's

lifetime. During early life, failure rates are higher and are classified as infant mortalities.

After approximately 100 hours of operation, the failure rate remains constant until the

device enters the wearout phase of its operating life. Wearout is rarely observed. Infant

mortality is attributed to manufacturing defects. Random defects account for failures

during normal life."</blockquote>

 

<p>Summary: Don't worry about it -- the shutter will fail long before the

semiconductors</p>

[gary_woodard]

somehow this is sad. What are we buying into.

[phule]

As Mike pointed out, there are pieces of the camera that will fail long long before the sensor. There are plenty of people shooting 100,000 photos a year with their digital slrs. You have spent far too much time and energy researching the wrong potential failure IMHO.

[thomas_sullivan]

kills time at work, so I research things....besides, it's electronics, so it's kinda work related ;o)

[thomas_sullivan]

btw - I still have not found an ACTUAL MTBF on CMOS sensors....................are they deliberately hiding them? is my first thought.

[phil vaughan - yorkshire u]

As above, I believe that as the CMOS is created from a sliver of silicon, it's quite stable, however the moving parts ie shutter, are going to wear out. How many PC's that are past their best are thrown out every day, these are full of electronic components that are working perfectly. However some of these PC's will be on their 2nd or 3rd hard drive, will have probably had more than one optical drive, and a couple of cooling fans.

 

Still they're not thrown away because they're expensive to maintain, but because they're out of date.

 

In 5 years time your 20d will maybe (MAYBE) have had a couple of shutter mechanisms, but you'll retire it because of a 600usd camera that is simply so much better, you can't resist.

[phule]

<< are they deliberately hiding them? is my first thought. >>

 

Yes. The camera companies don't want you to know that sensors will fail after just 2 days or 400 images, whichever comes first. They sneak in at night and swap out the sensor when you're asleep. If you don't sleep they have a team of commandos ready with knockout drops and the red flashy thing from MIB to erase your memory.

[thomas_sullivan]

heh.........point taken. It's just I work in electronics, and I know that specs on everything someone makes is available. I just wonder why this one is so hard to find.

 

So, instead of judging my reasons for even asking this question, is there anybody who has access to the actual info? And what about hot pixels?

 

By the way, you're responding to someone who just bought his first auto focus camera in the year 2000..........so, not an impluse buyer, and won't buy until he's happy with the answers. So, the last answer I even care about is the fact that it will be out of date in 3 years anyhow.........no offense meant, it's just how i view things.

[phil vaughan - yorkshire u]

Tom

 

No offence taken, I don't believe that the camera will be no good in 3 years, just that technology will have moved on enough in 5 years for even the most luddite to appreciate the difference in quality AND PRICE.

 

As for hot pixels and indeed dead pixels, all sensors have some, and they're mapped away in software by Canon, if you find new ones appearing, you can have them 'eliminated' by Canon.

[digitmstr]

Some NEW cameras may have stuck or dead pixels as well, though we hear of a few cases here and there.

 

Canon went from CCD to CMOS. Based on the CCD performance I don't think Canon would trade CCD for something that would have a *considerably* shorter life span. Especially since cameras like the 1DMKII are made for PJs who will be shooting far more pictures per year (in less than ideal conditions) than most amateurs. The 1D had a CCD, the 1DMKII has a CMOS. I doubt Canon would change for the worse.

 

If Canon doesn't publish figures about the CMOS they fabricate we can ONLY speculate, and that's all.

 

I would expect that every other part of the camera would fail before the CMOS...

 

I would say that on AVG most people with DSRL will take about 7,000 pix per year, or thereabouts. Because of the trend to keep buying new models it will be difficult to get a "personal experience" story on a CMOS failure. Of all my friends who own 10Ds none of them has complained of any CMOS issues and I think most amateurs will not keep their DSRL for 5 or more years.

[tom_burke3]

Actually the camera could be no good in 5 years time; there could be a failure of the

sensor, or the processing unit, or some other electronic component that may not be

economically repairable.

 

Basically, it comes down to this: we are all gambling on the lifetime of electronic

components. I assume that the manufacturers expect us to upgrade every two to three

years (just like most of us do with computers), and that the life of the electronic

components matches that expectation.

 

I also expect the major manufacturers to start offering official 'extended warranties' (if

they don't already) - Apple's 'AppleCare' is the sort of thing I'm thinking about, as a

model.

[jespdj]

MTBF for digital camera sensors is not an issue. Don't worry about it!

 

It's incredible how people in Internet forums worry about every potential problem that can possibly be conceived by the human imagination..... if you'd believe all that, you're never going to buy a digital camera because there's so much that potentially could be wrong with it!

[working with attitude sin]

I change computer each years, its my job. But i dont change my graphic tablet or monitor. Im sure people who are serious about photography change camera when a better come out but dont change their lens and tripod. Its life, you can have a taxes deduction on that kind of material.

[thomas_sullivan]

to those that actually answered my question.........thanks.

 

for the rest, it's not that I'm a luddite, I love new tech, think it's great. I also realize that things wear out, I'm 53 yo....I KNOW things wear out, believe me. What I don't buy into like some of the rest of you is that I should just believe the manufacturers marketting bullshit that I need a new camera every 3 years. Not only do they try to make me believe it with there BS, they apparently, from the comments above, make sure it will have the possibility to break down in that time period so I HAVE to buy a new one. That philosophy is total crap! And anybody that seriously buys into that deserves what they get.

 

I for one will ALWAYS research the products, of any substantial dollar figure, before I buy. I may still opt to buy it anyhow, but at least when I go into it, I will fully know what I am doing.

 

And that is all this post is about..............finding out information. When did it become stupid to want to seek out information when buying a product. I must have been busy taking pics on that day.

 

again, thanks to those that tried to help answer my questions.

[daniel_taylor]

Considering you can still buy working Commodore 64's and Apple II's from over 20 years ago, I doubt you have anything to worry about in terms of CCD/CMOS life. The two common things that can cause IC failure are high heat and poor power supply. DSLR's wouldn't get that hot even shooting in a desert at noon, and they've got their own clean, steady source of power (a battery). I suppose corrosion also plays a part, but that would almost certainly take many decades, or maybe even centuries.

 

The Canon AE-1 was one of the first SLR's to make heavy use of IC's in its design. Perfectly good, working models can be bought off ebay all day. Aren't these IC's between 25-30 years old?

 

I would be curious to see the MTBF for any chip, but I bet it's extreme across the board. Given a clean power supply and proper cooling there just really isn't any reason why an IC should fail in one person's lifetime. As has been pointed out repeatedly, the shutter will fail long, long before any of the IC's, as will the LCD backlighting.

[icephoto]

Okay, here's as close as I can come to an answer. Canon has a guarantee on the shutter mechanism on the 1D at approximately 150,000 shots. They also say that this is most likely the first mechanism to go (reference: Digital Photography Review interview with Canon http://www.dpreview.com/reviews/canoneos1d/page3.asp).

 

That's saying that the CMOS probably will last longer than that. But as an electronics tech you also know that Murphy worked on the 20D (as he has worked on every electronic and mechanical device over the last 2 millenia... ymmv).

 

But moving back to the concept of buying a new camera every 5-10 years... I look at digital cameras as an expendable rather than an investment. Why?

 

For every 36 images I shoot with the dSLR I've just saved $35-$60USD on film, processing, and prints and about 90 minutes of my time to do the film scans. I'm being conservative and say that my time is worth about $20 an hour, and in the worst case scenario if I need prints the following day, a roll of 36 with single prints and 24 hour developing will set me back about $30USD on the Icelandic economy. For those of you who bitch and moan when you get a few rolls done at your local pro lab and have to pull out your plastic, stop bitching.

 

At the best case scenario of $35 per roll, in about 1183 shots (about 33 rolls of 36) my 10D will have paid for itself. Now if I could be honest with myself I would dump $35 into a savings account each time I shot 36 images after it paid for itself then I would have a fund to buy the new one.

 

I don't believe in buying cameras just to buy cameras. I expect that my Elan 7e will last me another 10-15 years (or until film is discontinued) and that my 10D will last me equally as long. But I also see the reality of digital and when the 10D dies, I will replace it and will have very few tears.

 

If you want a rant, get me started on people who don't archive their digital images...

 

Shoot'em if you've got'em - jim

[charlie_vigue]

What I want to know is the MTBF of the microcontrollers in my microwave oven. I can't find that information anywhere, and I'm not buying a new one until I find this data! It's a conspiracy I tell you.

 

What? Door switches? Oh, nevermind then.

[jbq]

I remember the MTBF of CPUs being quoted as 15 years of continuous use. With CMOS sensors not being fundamentally different, it's safe to say that they could last a lifetime of continuous professional use.

 

Honestly, given the number of moving parts in a camera and the speed at which they're moving, the MTBF of the electronics is the last thing I'm worried about. The shutter and the mirror will likely have to be replaced many times before the electronics start to fail. 30000 exposures is what I would expect a shutter to last in a 10D/20D. Get a 1DII if you don't want to have to worry about the shutter failing.

[jeff medkeff anchorage, a]

Can't say anything about Canon's CMOS MTBF, but a friend has been building astronomical cameras with CMOS sensors recently, so I e-mailed him about this thread.

 

His answer - nobody really knows, but the manufacturer of his device guesstimates they are good for about sixty billion readouts or three hundred years, whichever comes first. I get the feeling these numbers have been pulled out of some engineer's darker regions, but....

 

This doesn't sound entirely unreasonable given other CMOS imaging sensor applications. How many readouts do video sensors live through? That's 29.97 readouts per second, for years of continuous service, for security cameras - 86,000 seconds per day for, say, five years (1825 days) comes out to 156,950,000 seconds which is 4,703,791,500 readouts. (Did I do that math right? Probably not, but I bet the numbers are still big.)

 

I'm told that readout amplifier life is a different story; you might only get ten or twenty billion readouts out of those devices.

 

Being mechanical, I bet the shutter has a fairly short lifetime, with failures falling on a gaussian distribution with really wide wings. So I'd expect some SLR owners to have shutter failures relatively early in the life of the camera.

[grant_.]

well, my only addition would be that the people ive been shooting with lately average at least a few GB of pix every single day....

<br>

<br>

thats probably 700-1000 images. multiply that by 5-7 for the days of the week, and figure out how many pix theyre getting from one sensor....the sensor if anything is the last thing to worry about... ;-)